CN107766373B - Method and system for determining categories to which pictures belong - Google Patents
Method and system for determining categories to which pictures belong Download PDFInfo
- Publication number
- CN107766373B CN107766373B CN201610694497.0A CN201610694497A CN107766373B CN 107766373 B CN107766373 B CN 107766373B CN 201610694497 A CN201610694497 A CN 201610694497A CN 107766373 B CN107766373 B CN 107766373B
- Authority
- CN
- China
- Prior art keywords
- category information
- picture
- retrieval
- query
- category
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/58—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/583—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
Landscapes
- Engineering & Computer Science (AREA)
- Library & Information Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The application discloses a method and a system for determining categories to which pictures belong, wherein the method comprises the following steps: receiving an input query picture; extracting picture features of the query picture, and determining a plurality of category information corresponding to the query picture according to the picture features; performing classified retrieval according to the query picture and the determined multiple category information; and judging the category information of the query picture according to the retrieval result of the classified retrieval. The accuracy of category inference can be improved through the method and the device.
Description
Technical Field
The present application relates to the field of computer technologies, and in particular, to a method and a system for determining categories to which pictures belong.
Background
Image retrieval has become increasingly common in e-commerce scenarios as an emerging technology. However, with the popularity of the internet, the number of pictures in the commodity picture library is increasingly huge, which makes image retrieval very difficult. Fortunately, if the commodity category to which the query picture (namely the user query picture in the picture search) belongs is known in advance, the vertical search can be carried out in the commodity category to which the query picture belongs, so that the search efficiency is greatly improved; secondly, the consistency of the result image returned by the retrieval and the query image on the category can be ensured by limiting the category, and the user experience is improved. In addition, the category information of the query picture is very helpful for understanding the user behavior and analyzing the user intention. Therefore, the category inference of the query picture has important significance for image searching.
Most of the category identification methods in the prior art are based on a learning strategy, a category confidence coefficient is output for each category, and the category with the highest confidence coefficient is usually selected as a prediction category for category identification. However, in a real e-commerce scene, the categories are of many kinds, and there are high apparent similarities among different categories (such as top dress and skirt dress), so that the confidence of category prediction of a partial query picture is not high, and the accuracy of category identification is limited (for example, the accuracy of the category with the highest confidence is only 80%). That is, there are many query pictures that result in inaccurate final search results due to wrong category prediction, and therefore, it is necessary to provide improved technical means to solve the above problems.
Disclosure of Invention
The present application mainly aims to provide a method and a system for determining categories to which pictures belong, so as to solve the problem in the prior art that a retrieval result obtained by querying a picture is inaccurate.
In order to solve the above problem, according to an embodiment of the present application, a method for determining a category to which a picture belongs is provided, including: receiving an input query picture; extracting picture features of the query picture, and determining a plurality of category information corresponding to the query picture according to the picture features; performing classified retrieval according to the query picture and the determined multiple category information; and judging the category information of the query picture according to the retrieval result of the classified retrieval.
The step of determining a plurality of category information corresponding to the query picture according to the picture features includes: and identifying the picture characteristics of the query picture according to a preset classifier, and calculating the attribution probability corresponding to each category information in a plurality of category information corresponding to the query picture.
Wherein, before the step of performing a classified search, the method further comprises: judging whether the maximum attribution probability in the attribution probabilities corresponding to each category information obtained through calculation is larger than a preset threshold value or not, if so, judging that the category information corresponding to the attribution probability is the category information to which the query picture belongs; otherwise, executing the step of classified retrieval.
Wherein the method further comprises: sequencing each category information according to the sequence of the attribution probability; and selecting part of category information in the plurality of category information according to the sequence for classified retrieval.
The step of determining the category information to which the query picture belongs according to the retrieval result of the classified retrieval includes: respectively acquiring a preset number of retrieval pictures corresponding to each category information; and respectively comparing the picture characteristics of the query picture with the picture characteristics of the retrieval picture, and if the picture characteristics of the query picture are the same, judging that the category information is the category information to which the query picture belongs.
Wherein if the picture characteristics of the query picture are determined to be different from the picture characteristics of the retrieval picture, the method further comprises: respectively acquiring a first quantity of retrieval pictures and a second quantity of retrieval pictures of each category information, wherein the first quantity is larger than the second quantity; respectively calculating the characteristic distances between the first number of retrieval pictures of each category information and the retrieval pictures, and determining the first category information with the minimum average distance; respectively calculating the characteristic distances between the retrieval pictures of the second quantity of each category information and the retrieval pictures, and determining the second category information with the minimum average distance; and if the first category information is the same as the second category information, judging that the category information is the category information to which the query picture belongs.
According to an embodiment of the present application, there is also provided a system for determining a category to which a picture belongs, including: the receiving module is used for receiving the input query picture; the category identification module is used for extracting the picture characteristics of the query picture and determining a plurality of category information corresponding to the query picture according to the picture characteristics; the multi-category retrieval module is used for carrying out classified retrieval according to the query picture and the determined multi-category information; and the category determining module is used for judging the category information to which the query picture belongs according to the retrieval result of the classified retrieval.
The category identification module is further configured to identify the picture features of the query picture according to a preset classifier, and calculate an attribution probability corresponding to each category information in the plurality of category information corresponding to the query picture.
Wherein, still include: the judging module is used for judging whether the maximum attribution probability in the attribution probabilities corresponding to each category information obtained through calculation is larger than a preset threshold value or not, and if yes, judging that the category information corresponding to the attribution probability is the category information to which the query picture belongs; otherwise, executing the step of classified retrieval.
The multi-category retrieval module is further used for sequencing each category information according to the sequence of the attribution probability; and selecting part of category information in the plurality of category information according to the sequence for classified retrieval.
The category determining module is further configured to obtain a preset number of retrieval pictures corresponding to each category information; and respectively comparing the picture characteristics of the query picture with the picture characteristics of the retrieval picture, and if the picture characteristics of the query picture are the same, judging that the category information is the category information to which the query picture belongs.
The category determining module is further configured to, if the category information is determined to be different, respectively obtain a first number of search pictures and a second number of search pictures of each category information, where the first number is greater than the second number; respectively calculating the characteristic distances between the first number of retrieval pictures of each category information and the retrieval pictures, and determining the first category information with the minimum average distance; respectively calculating the characteristic distances between the retrieval pictures of the second quantity of each category information and the retrieval pictures, and determining the second category information with the minimum average distance; and if the first category information is the same as the second category information, judging that the category information is the category information to which the query picture belongs.
According to the technical scheme of the application, when the confidence coefficient of the category inference based on deep learning is not high, classified retrieval is carried out according to the query picture and the determined plurality of category information, and therefore the category information to which the query picture belongs is determined. The accuracy of category inference can be improved through the method and the device.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:
FIG. 1 is a flow diagram of a method of determining a category to which a picture belongs according to one embodiment of the present application;
FIG. 2 is a flow chart of a method of determining a category to which a picture belongs according to another embodiment of the present application;
fig. 3 is a block diagram of a system for determining a category to which a picture belongs according to an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the technical solutions of the present application will be described in detail and completely with reference to the following specific embodiments of the present application and the accompanying drawings. It should be apparent that the described embodiments are only some of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
The technical solutions provided by the embodiments of the present application are described in detail below with reference to the accompanying drawings.
Fig. 1 is a flowchart of a method for determining a category to which a picture belongs according to an embodiment of the present application, as shown in fig. 1, the method includes:
step S102, receiving an input query picture (query picture).
When the user needs to perform image search, a query picture provided by a website or an image outside a website image database can be used as the query picture, for example, a picture taken by the user with a mobile phone, or a picture in another website, a local folder, or the like.
And step S104, extracting the picture characteristics of the query picture, and determining a plurality of category information corresponding to the query picture according to the picture characteristics.
In the embodiment of the application, a classifier based on deep learning is adopted for category inference, and category information corresponding to the query picture is determined. Specifically, firstly, extracting picture features of the query picture, such as CNN features; then, the image features of the query image are classified and identified according to a pre-trained CNN classifier, and the attribution probability of each item information in a plurality of item information corresponding to the query image, namely the item prediction confidence coefficient, is calculated.
Wherein the category prediction confidence of the category identification decision comes from the deep learning-based classifier. The classifier is trained in advance, a training model is established based on CNN, training data come from commodity pictures of an e-commerce platform, and each training data has identification information which indicates the category to which the training data belongs. And inputting the query picture into a classifier, outputting a vector corresponding to the query picture by the classifier, and representing the confidence coefficient of the query picture belonging to the corresponding category by the numerical value of each dimension of the vector. Generally, the category with the highest confidence level is taken as the prediction category, and the confidence level of the prediction category is the confidence level of the category.
In an embodiment of the present application, if the highest attribution probability is greater than a threshold, it is determined that the category information corresponding to the attribution probability is the category information to which the query picture belongs. Otherwise, step S106 is executed to perform multi-category retrieval analysis.
And S106, performing classified retrieval according to the query picture and the determined multiple category information.
In an embodiment of the application, each category information is sorted according to the order of the attribution probability, and part of the category information in the plurality of category information is selected according to the order for classified retrieval. For example, the top 3, 4 or 5 categories of the category prediction are selected according to the magnitude order of the attribution probability to be respectively queried, and the analysis and the category inference are carried out according to the top20 retrieval results (top20) returned by each category retrieval.
And S108, judging the category information of the query picture according to the retrieval result of the classified retrieval.
In one embodiment of the application, a preset number of retrieval pictures corresponding to each category information are respectively obtained; and respectively comparing the picture characteristics of the query picture with the picture characteristics of the retrieval picture, and if the picture characteristics of the query picture are the same, judging that the category information is the category information to which the query picture belongs.
In another embodiment of the present application, if the determinations are not the same, a quantitative statistical analysis is performed. Firstly, respectively acquiring a first quantity of retrieval pictures and a second quantity of retrieval pictures of each category information, wherein the first quantity is greater than the second quantity; then, respectively calculating the characteristic distances between the first number of retrieval pictures of each category information and the retrieval pictures, and determining the first category information with the minimum average distance; then, respectively calculating the characteristic distances between the retrieval pictures of the second quantity of each category information and the retrieval pictures, and determining the second category information with the minimum average distance; and if the first category information is the same as the second category information, judging that the category information is the category information to which the query picture belongs.
Details of the above process are described in detail below in conjunction with fig. 2. Fig. 2 is a flowchart of a method for determining a category to which a picture belongs according to another embodiment of the present application, as shown in fig. 2, the method includes:
in step S202, an input query picture (query picture) is received.
And step S204, extracting the picture characteristics of the query picture. Wherein the picture features are features capable of representing the essence of an image, including but not limited to: CNN features, SIFT features, SURF features, color features, texture histogram features. The CNN feature will be described as an example.
Step S206, determining a plurality of category information corresponding to the query picture according to the picture characteristics.
In the embodiment of the application, a CNN (Convolutional Neural Network) classifier based on deep learning is adopted to perform category inference, and category information corresponding to the query picture is determined. Specifically, firstly, extracting the CNN characteristics of the query picture; then, the image features of the query image are classified and identified according to a pre-trained CNN classifier, and the attribution probability of each item information in a plurality of item information corresponding to the query image, namely the item prediction confidence coefficient, is calculated.
For example, for query0, the output vector is [ 0.30.60.1 ]. That is, there are three categories, which in turn are, according to the correspondence: and if the query picture belongs to clothes, shoes and furniture, the confidence coefficient of the query picture is 0.3, the confidence coefficient of the shoes is 0.6 and the confidence coefficient of the furniture is 0.1. With the highest confidence for the shoe, query0 would typically be considered to belong to the shoe, with a category confidence of 0.6.
Step S208, determining whether the maximum attribution probability of the attribution probabilities corresponding to each category information is greater than a preset threshold, if so, performing step S218, otherwise, performing step S210.
And step S210, performing multi-category classified retrieval according to the query picture and the determined multi-category information. For example, the top 3, 4 or 5 categories of the category prediction are selected according to the magnitude order of the attribution probability to be respectively queried.
Step S212, obtaining a preset number of retrieval pictures corresponding to each category information.
For example, the top20 search results (top20) returned by each category search are obtained for analysis and category inference.
In step S214, it is determined whether the picture characteristics of the query picture are respectively the same as the picture characteristics of the retrieved picture, if so, step S218 is performed, otherwise, step S216 is performed.
For example, in the first 20 search results (top20) returned by each category, whether the search results are identified as the same graph or the same type as the query picture respectively, and if the search results are determined as the same graph or the same type, the category is considered to be the category information to which the query picture belongs.
Step S216, carrying out quantitative statistical analysis on the retrieval result of each category.
For example, respectively calculating the characteristic distances between the first 20 search results of each category and the query picture, and finding the category with the minimum average distance, which is denoted as C20; calculating the characteristic distance between the first 3 search results of each category and the query picture, and finding the category with the minimum average distance, which is marked as C3; if C20 and C3 are consistent, the category is considered as the category information to which the query picture belongs.
Step S218, determining that the category information is used as the category information to which the query picture belongs.
Fig. 3 is a block diagram of a structure of a system for determining a category to which a picture belongs according to an embodiment of the present application, and as shown in fig. 3, the system includes:
the receiving module 310 is configured to receive an input query picture.
The category identification module 320 is configured to extract picture features of the query picture, and determine a plurality of category information corresponding to the query picture according to the picture features. Specifically, the category identification module 320 identifies the picture features of the query picture according to a preset classifier, and calculates an attribution probability corresponding to each category information in the plurality of category information corresponding to the query picture.
And the multi-category retrieval module 330 is configured to perform category retrieval according to the query picture and the determined multiple categories of information. Specifically, the multi-category retrieving module 330 is further configured to sort each category information according to the order of the attribution probability; and selecting part of category information in the plurality of category information according to the sequence for classified retrieval.
And the category determining module 340 is configured to determine, according to the retrieval result of the classified retrieval, category information to which the query picture belongs.
The system further comprises: a determining module (not shown) configured to determine whether a maximum attribution probability of the attribution probabilities corresponding to each piece of category information obtained through calculation is greater than a preset threshold, and if so, determine that the category information corresponding to the attribution probability is the category information to which the query picture belongs; otherwise, executing the step of classified retrieval.
In an embodiment of the present application, the category determining module 340 is further configured to obtain a preset number of retrieval pictures corresponding to each category information; comparing the picture characteristics of the query picture with the picture characteristics of the retrieval picture respectively, and if the picture characteristics of the query picture are the same, judging that the category information is the category information to which the query picture belongs; if the retrieval images are judged to be different, respectively acquiring a first quantity of retrieval images and a second quantity of retrieval images of each category information, wherein the first quantity is larger than the second quantity; respectively calculating the characteristic distances between the first number of retrieval pictures of each category information and the retrieval pictures, and determining the first category information with the minimum average distance; respectively calculating the characteristic distances between the retrieval pictures of the second quantity of each category information and the retrieval pictures, and determining the second category information with the minimum average distance; and if the first category information is the same as the second category information, judging that the category information is the category information to which the query picture belongs.
The operation steps of the method correspond to the structural features of the system, and can be referred to one another, which is not described in detail.
As described above, the above embodiments according to the present application can achieve the following effects:
(1) the search range is narrowed, and the retrieval efficiency is greatly improved;
(2) by limiting the categories, the consistency of the result image returned by the retrieval and the query image on the categories can be ensured, and the user experience is effectively improved;
(3) the category information of the query picture is beneficial to understanding the user behavior and analyzing the user intention.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
The memory may include forms of volatile memory in a computer readable medium, Random Access Memory (RAM) and/or non-volatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of a computer-readable medium.
Computer-readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), Digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The above description is only an example of the present application and is not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.
Claims (12)
1. A method for determining a category to which a picture belongs is characterized by comprising the following steps:
receiving an input query picture;
extracting picture features of the query picture, and determining a plurality of category information corresponding to the query picture according to the picture features;
performing classified retrieval according to the query picture and the determined multiple category information;
judging the category information of the query picture according to the retrieval result of the classified retrieval;
and if the picture characteristics of the query picture are different from those of the retrieval pictures, determining the category information to which the query picture belongs according to the query picture, a first number of retrieval pictures of the category information in the plurality of category information and a second number of retrieval pictures, wherein the first number is larger than the second number.
2. The method according to claim 1, wherein the step of determining the category information corresponding to the query picture according to the picture feature comprises:
and identifying the picture characteristics of the query picture according to a preset classifier, and calculating the attribution probability corresponding to each category information in a plurality of category information corresponding to the query picture.
3. The method of claim 2, wherein prior to the step of performing a classification search, the method further comprises:
judging whether the maximum attribution probability in the attribution probabilities corresponding to each category information obtained through calculation is larger than a preset threshold value or not, if so, judging that the category information corresponding to the attribution probability is the category information to which the query picture belongs; otherwise, executing the step of classified retrieval.
4. The method of claim 2, further comprising:
sequencing each category information according to the sequence of the attribution probability;
and selecting part of category information in the plurality of category information according to the sequence for classified retrieval.
5. The method according to claim 1, wherein the step of determining category information to which the query picture belongs according to the search result of the classified search comprises:
respectively acquiring a preset number of retrieval pictures corresponding to each category information;
and respectively comparing the picture characteristics of the query picture with the picture characteristics of the retrieval pictures, and if the picture characteristics of the query picture and the picture characteristics of any retrieval picture are judged to be the same, judging that the category information is the category information to which the query picture belongs.
6. The method according to claim 5, wherein if it is determined that the picture characteristic of the query picture is not the same as the picture characteristic of the search picture, determining the category information to which the query picture belongs according to the query picture, the first number of search pictures and the second number of search pictures of the category information in the plurality of category information comprises:
respectively acquiring a first quantity of retrieval pictures and a second quantity of retrieval pictures of each category information;
respectively calculating the characteristic distances between the first number of retrieval pictures of each category information and the query picture, and determining the first category information with the minimum average distance;
respectively calculating the characteristic distances between the retrieval pictures of the second quantity of each category information and the query picture, and determining the second category information with the minimum average distance;
and if the first category information is the same as the second category information, judging that the category information is the category information to which the query picture belongs.
7. A system for determining a category to which a picture belongs, comprising:
the receiving module is used for receiving the input query picture;
the category identification module is used for extracting the picture characteristics of the query picture and determining a plurality of category information corresponding to the query picture according to the picture characteristics;
the multi-category retrieval module is used for carrying out classified retrieval according to the query picture and the determined multi-category information;
the category determining module is used for judging the category information to which the query picture belongs according to the retrieval result of the classified retrieval;
and the determining module is used for determining the category information to which the query picture belongs according to the query picture, a first quantity of retrieval pictures of category information in the plurality of category information and a second quantity of retrieval pictures if the picture characteristics of the query picture are judged to be different from the picture characteristics of the retrieval pictures, wherein the first quantity is larger than the second quantity.
8. The system according to claim 7, wherein the category identification module is further configured to identify picture features of the query picture according to a preset classifier, and calculate an attribution probability corresponding to each category information in a plurality of category information corresponding to the query picture.
9. The system of claim 8, further comprising:
the judging module is used for judging whether the maximum attribution probability in the attribution probabilities corresponding to each category information obtained through calculation is larger than a preset threshold value or not, and if yes, judging that the category information corresponding to the attribution probability is the category information to which the query picture belongs; otherwise, executing the step of classified retrieval.
10. The system of claim 8,
the multi-category retrieval module is further used for sequencing each category information according to the sequence of the attribution probability; and selecting part of category information in the plurality of category information according to the sequence for classified retrieval.
11. The system of claim 7, wherein the category determination module is further configured to,
respectively acquiring a preset number of retrieval pictures corresponding to each category information;
and respectively comparing the picture characteristics of the query picture with the picture characteristics of the retrieval pictures, and if the picture characteristics of the query picture and the picture characteristics of any retrieval picture are judged to be the same, judging that the category information is the category information to which the query picture belongs.
12. The system of claim 11, wherein the determination module is further configured to,
respectively acquiring a first quantity of retrieval pictures and a second quantity of retrieval pictures of each category information;
respectively calculating the characteristic distances between the first number of retrieval pictures of each category information and the query picture, and determining the first category information with the minimum average distance;
respectively calculating the characteristic distances between the retrieval pictures of the second quantity of each category information and the query picture, and determining the second category information with the minimum average distance;
and if the first category information is the same as the second category information, judging that the category information is the category information to which the query picture belongs.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610694497.0A CN107766373B (en) | 2016-08-19 | 2016-08-19 | Method and system for determining categories to which pictures belong |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610694497.0A CN107766373B (en) | 2016-08-19 | 2016-08-19 | Method and system for determining categories to which pictures belong |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107766373A CN107766373A (en) | 2018-03-06 |
| CN107766373B true CN107766373B (en) | 2021-07-20 |
Family
ID=61262615
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610694497.0A Active CN107766373B (en) | 2016-08-19 | 2016-08-19 | Method and system for determining categories to which pictures belong |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107766373B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109598286A (en) * | 2018-10-25 | 2019-04-09 | 努比亚技术有限公司 | A kind of picture classification method, terminal and computer readable storage medium |
| CN111833298B (en) * | 2020-06-04 | 2022-08-19 | 石家庄喜高科技有限责任公司 | Skeletal development grade detection method and terminal equipment |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101196905A (en) * | 2007-12-05 | 2008-06-11 | 覃征 | Intelligent pattern searching method |
| CN101551809A (en) * | 2009-05-13 | 2009-10-07 | 西安电子科技大学 | Search method of SAR images classified based on Gauss hybrid model |
| CN102057371A (en) * | 2008-06-06 | 2011-05-11 | 汤姆逊许可证公司 | System and method for similarity search of images |
| CN103440262A (en) * | 2013-07-31 | 2013-12-11 | 东莞中山大学研究院 | Image searching system and image searching method basing on relevance feedback and Bag-of-Features |
| CN104346370A (en) * | 2013-07-31 | 2015-02-11 | 阿里巴巴集团控股有限公司 | Method and device for image searching and image text information acquiring |
| CN105354307A (en) * | 2015-11-06 | 2016-02-24 | 腾讯科技（深圳）有限公司 | Image content identification method and apparatus |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2402867B1 (en) * | 2010-07-02 | 2018-08-22 | Accenture Global Services Limited | A computer-implemented method, a computer program product and a computer system for image processing |
| CN103679188A (en) * | 2012-09-12 | 2014-03-26 | 富士通株式会社 | Image classifier generating method and device as well as image classifying method and device |
| CN103714353B (en) * | 2014-01-09 | 2016-11-23 | 西安电子科技大学 | The Classification of Polarimetric SAR Image method of view-based access control model prior model |
| CN103955709B (en) * | 2014-05-13 | 2017-04-19 | 西安电子科技大学 | Weighted synthetic kernel and triple markov field (TMF) based polarimetric synthetic aperture radar (SAR) image classification method |
| CN105844283B (en) * | 2015-01-16 | 2019-06-07 | 阿里巴巴集团控股有限公司 | Method, image search method and the device of image classification ownership for identification |
-
2016
- 2016-08-19 CN CN201610694497.0A patent/CN107766373B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101196905A (en) * | 2007-12-05 | 2008-06-11 | 覃征 | Intelligent pattern searching method |
| CN102057371A (en) * | 2008-06-06 | 2011-05-11 | 汤姆逊许可证公司 | System and method for similarity search of images |
| CN101551809A (en) * | 2009-05-13 | 2009-10-07 | 西安电子科技大学 | Search method of SAR images classified based on Gauss hybrid model |
| CN103440262A (en) * | 2013-07-31 | 2013-12-11 | 东莞中山大学研究院 | Image searching system and image searching method basing on relevance feedback and Bag-of-Features |
| CN104346370A (en) * | 2013-07-31 | 2015-02-11 | 阿里巴巴集团控股有限公司 | Method and device for image searching and image text information acquiring |
| CN105354307A (en) * | 2015-11-06 | 2016-02-24 | 腾讯科技（深圳）有限公司 | Image content identification method and apparatus |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107766373A (en) | 2018-03-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106033416B (en) | Character string processing method and device | |
| CN110188223B (en) | Image processing method and device and computer equipment | |
| US7783581B2 (en) | Data learning system for identifying, learning apparatus, identifying apparatus and learning method | |
| CN102254015A (en) | Image retrieval method based on visual phrases | |
| US20150169998A1 (en) | Object detection in images based on affinity determinations | |
| CN114581207B (en) | Commodity image big data accurate pushing method and system for E-commerce platform | |
| CN111723226B (en) | Information management method based on big data and Internet and artificial intelligence cloud server | |
| US10489681B2 (en) | Method of clustering digital images, corresponding system, apparatus and computer program product | |
| CN110825894A (en) | Data index establishing method, data index retrieving method, data index establishing device, data index retrieving device, data index establishing equipment and storage medium | |
| CN109086830B (en) | Typical correlation analysis near-duplicate video detection method based on sample punishment | |
| CN107315984B (en) | Pedestrian retrieval method and device | |
| CN107766373B (en) | Method and system for determining categories to which pictures belong | |
| CN106407281B (en) | Image retrieval method and device | |
| CN112307199A (en) | Information identification method, data processing method, device and equipment, information interaction method | |
| CN111723227B (en) | Data analysis method based on artificial intelligence and Internet and cloud computing service platform | |
| CN112784008B (en) | Case similarity determining method and device, storage medium and terminal | |
| CN111738173A (en) | Video clip detection method and device, electronic equipment and storage medium | |
| CN114925239B (en) | Intelligent education target video big data retrieval method and system based on artificial intelligence | |
| CN115984671A (en) | Model online updating method and device, electronic equipment and readable storage medium | |
| CN108255880B (en) | Data processing method and device | |
| CN111984812B (en) | Feature extraction model generation method, image retrieval method, device and equipment | |
| CN109739840A (en) | Data processing empty value method, apparatus and terminal device | |
| CN113284027A (en) | Method for training group recognition model, and method and device for recognizing abnormal group | |
| CN112861974A (en) | Text classification method and device, electronic equipment and storage medium | |
| CN112488140A (en) | Data association method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |